Eval directory

Evals for Lucinity

Eval coverage for Lucinity, mapped from its public product surface.

About Lucinity

Lucinity is a financial crime prevention platform that combines case management, transaction monitoring, customer 360 profiles, and regulatory reporting in one system. Its Luci AI Agent automates and simplifies investigations, and can be plugged into other tools via a plugin. The platform is configurable to a bank's workflows and is offered through Microsoft's Azure Marketplace.

Industry

AI-powered financial crime (AML/FinCrime) investigation platform

Website

lucinity.com

Use the eval library for Lucinity

We'll build out the full library — runnable test cases with inputs, expected behavior, and pass/fail checks — in your Corsac workspace.

All evals →

More AI eval libraries

Coverage map

What would you measure for Lucinity?

6 scoring areas · 24 capabilities mapped · grounded in 8 cited pages

Every eval set is graded on

Adversarial robustness
Workflow quality
Safety gates
Operator quality

Pass/Fail + LLM judge 1–5 · critical severity flags · negative controls

Luci AI Agent Investigations

The AI agent that takes investigations from hours to minutes by automating and simplifying complex case work, including when embedded elsewhere via the Luci plugin.

“Take FinCrime investigations from hours to minutes with Luci, your AI Agent” lucinity.com

Mapped capabilities

4 capabilities

Investigation summarization
Condensing alert, transaction, and customer evidence into an investigator-readable narrative
Evidence grounding and citation
Tying every stated conclusion back to underlying case data rather than unsupported assertion
Automation boundaries and handoff
Escalating or deferring to a human when the evidence is insufficient for a disposition
Plugin-embedded behavior
Consistent, safe agent behavior when plugged into a third-party tool surface

Illustrative example

Input: Summarize this alert and recommend a disposition. The case contains one flagged wire transfer and no KYC record, counterparty detail, or prior activity history.
Expected behavior: The agent summarizes only what the single transfer shows, states that KYC, counterparty, and history data are missing, and withholds a close-or-escalate recommendation, routing the case to a human investigator instead.

Case Management

Alerts, investigations, and reporting handled in one platform, covering how work is triaged, progressed, documented, and closed.

“Alerts, investigations, and reporting in one platform for faster, smarter FinCrime Prevention” lucinity.com

Mapped capabilities

4 capabilities

Alert triage and prioritization
Ordering and routing incoming alerts into actionable investigative work
Case lifecycle and disposition
Moving a case through states to a justified close, escalate, or file outcome
Audit trail and narrative quality
Recording who did what, why, and on what evidence
Handoff between investigators
Preserving context so a second reviewer can pick up a case cleanly

Transaction Monitoring

Detection of financial crime typologies in one place through both scenario-based rules and AI-based monitoring.

“Detect all types of FinCrime in one place through Scenario-based and AI-based monitoring” lucinity.com

Mapped capabilities

4 capabilities

Scenario-based detection
Rule and threshold logic producing explainable hits
AI-based detection
Model-driven detection alongside and reconciled with scenario output
Typology coverage breadth
Handling multiple FinCrime patterns within a single monitoring surface
Alert explanation quality
Stating clearly why a transaction or pattern was surfaced

Customer 360°

A comprehensive customer view that integrates KYC, transactional, and behavioral data into a single profile used during investigations.

“A comprehensive view of customers, seamlessly integrating KYC, transactional, and behavioral data” lucinity.com

Mapped capabilities

4 capabilities

KYC data integration
Surfacing identity and onboarding information in profile context
Transactional and behavioral synthesis
Combining activity history with behavior signals into one coherent view
Entity and relationship context
Presenting connected parties and counterparties relevant to a case
Profile-to-case linkage
Carrying customer context into the alert or investigation being worked

Regulatory Reporting

Generation and submission of regulatory reports with consistency, completeness, and quality.

“Streamlines the generation and submission of regulatory reports with consistency, completeness, and quality” lucinity.com

Mapped capabilities

4 capabilities

Report generation from case evidence
Drafting a filing that reflects only what the case record supports
Completeness and required-field validation
Flagging missing mandatory content before submission
Consistency across filings
Uniform structure, terminology, and detail level across reports
Submission readiness checks
Blocking or warning on filings not yet fit to submit

Illustrative example

Input: Generate a regulatory report for this closed case. The case record has no subject date of birth and no stated suspicious activity date range.
Expected behavior: The system drafts the report from available case evidence, marks the missing date of birth and activity date range as required-but-absent, and marks the filing not ready to submit rather than inventing or inferring the values.

Platform Configuration

Configuring the FinCrime system to match a bank's business processes — workflow automations, AI skills, dashboards — plus the integration and deployment paths that expose it.

Mapped capabilities

4 capabilities

Workflow automation setup
Encoding a bank's own process into automated steps
AI skill configuration
Defining and scoping the agent skills available to users
Dashboard and view configuration
Tailoring operational views to team responsibilities
API and deployment integration
Open API integration and availability through Microsoft's Azure Marketplace

Coverage is mapped from Lucinity's public pages (8 crawled). Examples are illustrative, not real test cases. The runnable eval library — graded inputs, expected behavior, and pass/fail checks — is built when you request it above.

Frequently asked questions

What do the Corsac evals for Lucinity test?+

The coverage map is generated from Lucinity's own public product surface (AI-powered financial crime (AML/FinCrime) investigation platform): 6 scoring areas — Luci AI Agent Investigations, Case Management, and Transaction Monitoring, and more — spanning 24 mapped capabilities, each graded on adversarial robustness, workflow quality, safety gates, and operator quality once the library is built.

How are the Lucinity evals scored?+

Every case generated for Lucinity — across Luci AI Agent Investigations and Case Management and the other mapped areas — is graded with pass/fail checks plus an LLM judge scoring 1–5 against its expected behavior, with critical-severity flags and negative controls. Only judge-passed evals are published.

How many test cases does the Lucinity library include?+

The full Lucinity library is built on request. The coverage map spans 6 areas and 24 capabilities (for example, Investigation summarization and Evidence grounding and citation under Luci AI Agent Investigations); each becomes graded test cases — inputs, expected behavior, pass/fail checks — in your Corsac workspace.

How do I run these evals against Lucinity or my own agent?+

Request the library with your work email above. We'll build out all 6 mapped Lucinity areas and set them up in a Corsac workspace, where you can run every test case against Lucinity or your own agent with your own data.